EP0025397A1 - Illuminating means for producing a light beam having adjustable light intensity distribution and image transfer system comprising such a means - Google Patents

Abstract

The invention relates to an illuminating device comprising a laser source providing a coherent beam with Gaussian intensity distribution and means for separating (5,6) the initial beam into four parts (F11-F22) in two orthogonal directions (x, y ), means of recombination (Mx1 - My2) of the four parts leading to their partial superposition in a predetermined plane (P) and means of averaging (L, K1, K2) of the interference produced in the plane (P) by the four parts. Application in particular to direct photo-repetition.

Description

The present invention relates to an illuminating device intended to provide a quasi-coherent illumination beam having a spatial distribution of intensity adjustable at will.

A laser beam generally has a Gaussian intensity distribution with circular symmetry along the section due to the fact that only the fundamental mode resonates in the cavity. In an illuminating device intended for the illumination of an object, for example in a pattern projection system, the beam coming from the laser is widened and made divergent. It retains its Gaussian distribution, so that the center of the object is lit more intensely than the edges. However, to obtain good image quality and also identical line widths throughout the field of the projection objective, in particular when the patterns are projected onto a photosensitive resin, good uniformity of illumination in the plane is required. picture. Thus, in the case of the production of integrated circuits by direct photo-repetition, which involves the formation of a superposition of levels in a semiconductor substrate, with successive masks, the thinnest lines have a width of the order of micron. The line width dispersion must be less than 10%, which implies variations in light intensity less than 3%.

Various methods have been proposed for making uniform illumination: the simplest method consists in broadening the beam sufficiently to use only the center spokes. This method, to be effective, has an energy yield of the order of 1%, which removes much of the advantage of using a laser. Other methods generally using beam scattering or absorption also have very low light output. However, some methods do not have this drawback. We can use a digital hologram plotted by computer and approximating by a binary function the complex transformation giving a uniform spatial distribution from a Gaussian distribution. This method is complex and costly to implement. We can also take into account the fact that the Gaussian distribution is due to the fact that the laser is single-mode, and use for the laser mirrors of higher radius of curvature, which allows other modes to resonate. An intensity distribution is obtained with a less pronounced maximum, but the lighting cannot be standardized sufficiently by this means.

The present invention makes it possible to obtain uniform illumination within less than 5%, without loss of energy, and with simple optical elements. In addition, the device according to the invention makes it possible to adjust the intensity distribution, within certain limits, to obtain non-uniform illuminations, but corresponding to the best illumination of the image plane, as a function of the defects of the optical projection system.

According to the invention, the illuminating device comprises means allowing the separation of the initial beam into four parts, in two radial directions perpendicular to each other, means allowing the recombination of the four parts culminating in a predetermined plane (P) partial of the four beams and the means for averaging the the interferences between the four parts so that the average intensity in the plane P is equal to the sum of the average intensities of the four parts.

• - la figure 1 montre la répartition d'intensité d'un faisceau laser ;
• - la figure 2 représente schématiquement un dispositif illuminateur ;
• - la figure 3 représente schématiquement un dispositif de projection de motifs ;
• - la figure 4 illustre les répartitions d'intensité dans les plans objet et image ;
• - la figure 5 illustre schématiquement la méthode utilisée selon l'invention ;
• - la figure 6 représente un mode de réalisation de l'invention ;
• - les figures 7 et 8 sont des vues partielles détaillées ;
• - la figure 9 représente une variante de réalisation de l'invention ;
• - la figure 10 montre le résultat obtenu par le dispositif de la figure 6 ou de la figure 9 :
• - la figure 11 représente une application de l'invention ;
  La figure 1 montre la répartition d'intensité d'un faisceau laser monomode. Un tel faisceau se caractérise essentiellement par sa très faible divergence (de l'ordre du milliradian) et une répartition d'intensité gaussienne suivant la section du faisceau. Ainsi, les rayons proches du centre du faisceau sont intenses, l'intensité sur l'axe central ayant-une valeur I₀; tandis que les rayons des bords du faisceau sont de faible intensité. Les bords du faisceau peuvent être définis par leur distance radiale W, visible sur la figure 1a que l'on appelera rayon du faisceau telle que l'intensité soit égale à 0_. La figure 1b montre que la distribution est a symétrie circulaire suivant la section du faisceau, constituée par le plan (x, y).

Other characteristics and advantages of the invention will become clear with the aid of the description below and the appended figures among which:

• - Figure 1 shows the intensity distribution of a laser beam;
• - Figure 2 schematically shows an illuminating device;
• - Figure 3 schematically shows a pattern projection device;
• - Figure 4 illustrates the intensity distributions in the object and image planes;
• - Figure 5 schematically illustrates the method used according to the invention;
• - Figure 6 shows an embodiment of the invention;
• - Figures 7 and 8 are detailed partial views;
• - Figure 9 shows an alternative embodiment of the invention;
• - Figure 10 shows the result obtained by the device of Figure 6 or Figure 9:
• - Figure 11 shows an application of the invention;
  Figure 1 shows the intensity distribution of a single mode laser beam. Such a beam is essentially characterized by its very small divergence (of the order of a milliradian) and a distribution of Gaussian intensity according to the section of the beam. Thus, the rays close to the center of the beam are intense, the intensity on the central axis having a value I ₀ ; while the rays of the edges of the beam are of weak intensity. The edges of the beam can be defined by their radial distance W, visible in FIG. 1a which will be called beam radius such that the intensity is equal to 0 _. Figure 1b shows that the distribution is circularly symmetrical along the section of the beam, formed by the plane (x, y).

A laser beam can be used in an illuminating device such as that, shown by way of example in FIG. 2. The beam F delivered by the laser 1 and having the intensity distribution. Represented in FIG. 1, is widened by an expander formed by lenses L ₁ and L ₂ . It is made divergent by a lens L ₃ and condensed by a condenser 3 so as to illuminate an object 4. The intensity distribution in the plane of the object remains Gaussian, so that the edges of the object are less illuminated than the center. However, depending on the applications, it is desired to obtain a predetermined intensity distribution, adapted to the application.

One of the possible applications is represented in FIG. 3. It is a projection device comprising in particular a lens 9 forming the image of the object 4 in a plane 12. To obtain good image quality in all the field of the objective, it is necessary a good uniformity of illumination of the image plane.

entre l'intensité au bord et au centre, doit être de 5 % à 10%.FIG. 4 shows, in two cases, the intensity distributions obtained in the image plane. In the first case, the distribution in the object plane, represented in FIG. 4a, is uniform. It is then generally observed that the illumination in the image plane shown in FIG. 4b is non-uniform, more intense in the center and gradually decreasing towards the edge of the lens field. This is due to the light losses in the different lenses which form the objective 9 and which are generally larger at the edges than in the center, because of the spherical shape of the lenses. Consequently, to have uniform illumination in the image plane, as shown in FIG. 4c, less intense in the center than at the edges. Typically, the maximum deviation

between the intensity at the edge and in the center, should be 5% to 10%.

FIG. 5 schematically illustrates the method used according to the invention for obtaining an illumination of the type of that of FIG. 4c. We will see below that the invention also makes it possible to obtain other forms of lighting close to uniform lighting.

The original beam F has in any radial direction such that x, a Gaussian intensity distribution shown in FIG. 5a. The invention provides for separating this beam into two half-beams along an axial plane perpendicular to the x axis, then shifting the two beams obtained by the same quantity a, in the opposite direction. The two half-beams are superimposed, as shown in FIG. 5b, and a resulting beam is obtained, the intensity distribution of which is represented in FIG. 3c, provided that the two half-beams are added in intensity, and not in phase. and in amplitude, which can be obtained for example, by rotating the polarization of one of the half-beams by 90 °. We note that we end up with a quasi-uniform distribution along the direction x. The calculations show that, for a = 0.555 W, the maximum intensity difference is 3%. On the other hand, the distribution remains Gaussian along the radial axis y perpendicular to x. It is possible to perform the same separation-recombination along the y axis, with the same value for the offset. Assuming that we have made the beams add in intensity in the y direction as well (the rest of the description will show how this is possible), we show that the shape of the beam approaches a square, which is often advantageous because the useful lighting fields of objects are often square, hence a gain in light output. On the other hand, the intensity distribution is well defined: weaker in the center, it gradually increases from I ₀ to 103% I ₀ towards the edge along the x and y axes and from I ₀ to 106% I ₀ towards the edge along the diagonals. Such a beam therefore provides an illumination close to the ideal illumination, shown in FIG. 4c.

FIG. 6 represents an optical assembly ensuring the separation-recombination of a laser beam in two orthogonal directions, for the realization of the invention. The laser beam F delivered by the laser 1 is widened by an expander 2. The separation along an x axis is ensured by a Fresnel biprism 5 suitably oriented and centered with respect to the beam F. The beam F is thus separated into two half bundles F ₁ and F ₂ . To simplify the figure, only the rays resulting from the central radius of the beam F, which forms the optical axis z of the system, have been shown. The beams F ₁ and F ₂ are reflected respectively by two mirrors M _x1 and M _x2 , so as to make them almost parallel to the axis z. They then cross a second Fresnel biprism 6, oriented orthogonal to biprism 5 and which ensures the separation of each beam F ₁ and F ₂ along the axis y perpendicular to x, respectively forming the beams F ₁₁ and F _{12 '} F ₂₁ and F ₂₂ . The beams F ₁₁ and F ₂₁ are reflected by a mirror M _y1 and the beams F1 ₂ and F ₂₂ are reflected by a mirror M _y2 , so as to make them again almost parallel to the axis z. The mirrors M _y1 and P _y2 are placed so that the respective impact points A, B, C, D of the beams F ₁₂ , F _{11 '} F _21' F ₂₂ are seen at the same inclination x relative to the z axis, in the recombination plane P of the beams. It is in this plane that the first lens L ₁ of the device can be placed of illumination of FIG. 2. To preserve the overall collimating properties of the two lenses L ₁ and L ₂ of FIG. 2, the angle x must be very small.

The operation of the first separator constituted by the biprism 5 and the mirrors M _x1 and M _x2 can be explained from the detailed view of FIG. 7. The beams obtained F ₁ and F ₂₁ have respectively, along the axes x ₁ and x ₂ perpendicular to their direction of propagation, intensity distributions in the form of a semi-Gaussian, as illustrated in FIG. 5.

FIG. 8 shows the path of the beams F _{11 '} F _12' F _{21 '} F ₂₂ at the outlet of the second separator, formed for biprism 6 and the mirrors M _y1 and My ₂ . The four points A, B, C, D form a square with side 2b in an x, y plane perpendicular to the z axis centered on the z axis. The rays of the beams F _{11 '} F _21' F _{12 '} F ₂₂ coming from the central radius of the initial beam F form the same angle x with the axis Z. These four beams form in the plane P an interference system because the four sources are consistent with each other because of the coherence of the laser. As the four beams are not parallel to the optical axis, orthogonal fringes are observed along the axes x _p and y _p of the plane P, forming a regular grid. By acting on the polarization of the beams, one can ensure that one of the two fringe systems is eliminated. For example, if a half-wave plate L is interposed on the beam path F _21, the beam F being already linearly polarized, it is possible to ensure that the beams F ₂₂ and F ₂₁ on the one hand and F ₁₁ and F ₁₂ on the other hand are respectively polarized in directions P ₂ and P ₁ perpendicular to each other, and therefore no longer interfere. However, these beams interfere with each other and form two systems of fringes of the same direction, independent and which overlap. For these two systems, it is necessary to know the laws of phase distribution of the four beams, according to the laws of phase of different optical elements crossed: the half-wave plate and the two biprisms, the reflection on the mirrors not modifying the distributions phase, and also, the modification observed in the plane P due to the inclination x of the beams, which depends on the value of the angle a, of b and on the distance h between the plane P and the square ABCD. We can then calculate the respective phase shifts of the four beams. If a first electro-optical crystal K is introduced into the path of one of the beams F ₁₁ and F ₁₂ , polarized by an electric field parallel to the direction of polarization common to these two beams and whose value varies in the time, one can introduce an additional phase shift variable in time and one can make so that the average value of cos ϕ on a duration T low compared to the duration envisaged so that the illumination is null. The calculation then shows that the average value of the resulting intensity is equal to the sum of the intensities of each of the two beams. Similarly, it is possible to introduce into the path of one of the two beams F ₂₁ and F ₂₂ a second electro-optical crystal K ₂ polarized by an electric field parallel to the direction of polarization common to these two beams and the value of which varies in time, can be the same reasoning and show that when the average value _of con- era, the two beams F ₂₁ and F ₂₂ are added in intensity and not in amplitude and phase. Like the beams F ₁₁ and F ₁₂ on the one hand, ^F ₂₁ and ^F ₂₂ on the other hand add in intensity; because of their perpendicular polarizations, one obtained, not exactly a suppression of the interferences, but an averaging in time, over a sufficiently short duration compared to the duration of illumination or observation so that the illumination can be considered homogeneous, for example as to its effects on the printing of a photosensitive resin.

Le cristal K₂ est placé sur le trajet du faisceau F₂₂ et il est commandé par un champ E₂ parallèle à la polarisation de ce faisceau, donc à l'axe x. Il génère un déphasage φ₂, qui peut être de la forme

où φ₀ a une valeur fixe quelconque.The half-wave plate L and the electro-optical crystals K ₁ and K ₂ are shown in FIG. 6, as well as the directions of polarization of the beams, symbolized by arrows. The beam F is polarized parallel to the y axis. The half-wave plate L is placed on the path of the beam L in an xy plane. Its neutral lines are oriented at 45 ° from the x, y axes, so that, at the exit of the blade, the polarization of the beam F ₂ is parallel to the x axis. The crystal K ₁ is placed on the path of the beam F ₁₂ and it is controlled by a field E _l parallel to the polarization of this beam, therefore to the y axis. It generates a phase shift φ ₁ , which can for example be equal to

The crystal K ₂ is placed on the path of the beam F ₂₂ and it is controlled by a field E ₂ parallel to the polarization of this beam, therefore to the x axis. It generates a phase shift φ ₂ , which can be of the form

where φ ₀ has any fixed value.

Les directions de polarisation des deux faisceaux restent inchangés.The same interference averaging result is obtained by reversing the two means used: half-wave plate and phase shifters. FIG. 9 shows the same optical arrangement for separation-recombination of the beam, only the means making it possible to average the interference being different. The beam F being always polarized along the y axis, a phase shifter is introduced into the path of the beam F ₂ constituted by an electro-optical crystal K polarized by an electric field E parallel to y, introducing a phase shift φ between the beams F ₁ and F ₂ of the form

The polarization directions of the two beams remain unchanged.

A half-wave plate L ₁ is introduced on the path of the beam F ₁₂ and a half-wave plate L ₂ is introduced on the path of the beam _F22 , so that the polarizations of these two beams become parallel to the x axis. Thus, in plane P, the two fringe systems are eliminated. This method is less expensive than the previous one because it uses only one electro-optical phase shifter instead of two.

Another possible method for eliminating the effect due to interference consists in placing on the path of the beams, a rotating diffusing element which generates time-varying, spatially random phase shifts. This method will be advantageously used if, instead of making the two separators separately with, for each separator, a biprism and two mirrors, one. performs the separations-recombinations respectively by means of a prism with circular symmetry, having a conical shape and a truncated-conical mirror. This is the generalization of the process which has just been described. At the exit of the conical prism the beam is transformed into an annular beam and recombined by the conical mirror. Such means are illustrated in Figure 11 which will be described later.

où T =

est faible devant la durée d'éclairement. In the case where the laser beam is not linearly polarized, the use of a birefringent plate is no longer possible unless interposing a polarizer on the path of the beam, which causes a loss of energy of 50 %. One can place on the path of the four beams four electro-optical phase shifters creating respective phase shifts, for example of the form:

where T =

is weak compared to the duration of illumination.

Whatever the means used on the one hand to ensure the separation-recombination of four beams, on the other hand to remove the effect of interference between the four beams, the separation and the recombination are illustrated in FIGS. 10a and 10b. The section of the beam F is shown in FIG. 10a. The device according to the invention breaks this beam into four sectors, then performs translations of type (± a, ± a). Thus, the beam F ₁₁ , hatched in the figure is translated according to the vector T (+ a, + a). In the plane P, represented in FIG. 10b, the four superimposed beams are inscribed in a square with side 2a centered on the axis z.

The value of a is determined by the distance h. If the plane P passes through the meeting point 0 of the four rays coming from the central radius of the beam F, there is no superposition, a = 0: the beam F is reconstituted. The more the distance h increases, the greater the superposition. It has been said previously that, when a is of the order of 0.5 W, a substantially uniform intensity distribution is obtained. For lower values of a, the intensity distribution is in the form of a "horse saddle", that is to say that the beam is intense on the edges and in the center, and less intense on intermediate places. For higher values of a, the distribution is bowl-shaped: more intense on the edges than in the center.

An illuminating device comprising means for uniformizing the beam is applicable in particular to a photorepeater.

FIG. 11 represents a photorepeater of the type described in French patent application no. 76.37.327 and published under no. 2 406 236. Behind the laser 1, for example a Krypton laser of power 1.5 W, are arranged the beam expander 2, a first deflection mirror R ₁ and the uniformization means, constituted by a conical prism C, a T-shaped element cone section with reflecting walls and a rotating diffuser D. The beam thus having the desired intensity distribution is returned by a second mirror R ₂ to a shutter 14 controlled so that the amount of energy received by the layer of photosensitive resin intended to be impressed corresponds to the best possible sunshine. The control means described in the first cited request, are mainly constituted by an integrator 15, an energy quantity selector 17 and a comparator device 16 triggering the opening and closing of the shutter 14. They are then placed on the path of the beam of the means 200 intended to produce a quasi-point source of dimensions such that no disturbing interference patterns occur.

The divergent beam 100 from the means 200 illuminates via a condenser 3 the object 4 carried by a holder 8 and whose image is formed by the objective 9 carried by a objective 10 provided with automatic focusing means 11 on the substrate 12 covered with a photosensitive resin and carried by a movable table 13. The illumination obtained thanks to the conical prism C and to the frusto-conical mirror T allows, as has been shown above, to obtain a good quality of image at the level of the photosensitive resin.

Claims (9)

1. Illuminating device intended to provide a coherent illumination beam having a spatial distribution of adjustable intensity, from a source (1) providing an initial beam (F) having a Gaussian intensity distribution (I), characterized in that it comprises means (5 - 6) allowing the separation of the initial beam into four parts each forming a beam (F ₁₁ - F22), in two radial directions (x, y) perpendicular to each other, means (M _x1 - M _y2 ) allowing the recombination of the four parts leading in a predetermined plane (P) to a partial superposition of the four beams (F ₁₁ - F ₂₂ ) and means (L, K ₁ - K ₂ ) allowing to average the interferences between the four parts so that the average intensity in the predetermined plane (P) is equal to the sum of the average intensities of the four parts. 2. Illuminating device according to claim 1, characterized in that the separation means comprise a first biprism (5) centered on the path of the initial beam (F) and separating it in a first radial direction x into first and second half beams (F ₁ - F ₂ ) symmetrical with respect to the axis (z) of the initial beam (F), and a second biprism (6) centered on the path of the two half-beams (F ₁ - F ₂ ) and the dividing respectively, in a second radial direction (y) perpendicular to the first, into first and second parts (F ₁₁ - F ₁₂ ) symmetrical with respect to the first half-beam (F1) and into third and fourth parts ( F ₂₁ - F ₂₂ ) symmetrical with respect to the second half-beam (F ₂ ). 3. Illuminating device according to claim 2, characterized in that the recombination means com take a first (M _x1 ) and a second (M _x2 ) mirrors respectively reflecting the two half-beams (F ₁ - F ₂ ) in two directions symmetrical with respect to the axis (z) of the initial beam (F) and converging towards the second biprism (6), and a third (M _y1 ) and a fourth (My ₂ ) mirrors reflecting respectively the first two parts (F ₁₁ F ₂₁ ) and the last two parts (F ₁₂ - F ₂₂ ) in four directions symmetrical about the axis (z) of the initial beam (z) and converging towards the predetermined plane (P). 4. Illuminating device according to claim 3, characterized in that, the initial beam (F) being linearly polarized, the averaging means comprise at least one birefringent plate (L) located on the path of one of the half-beams ( F ₂ ) so that these two half-beams have directions of polarization perpendicular to each other, and at least two electro-optical phase shifting elements (K ₁ - K ₂ ) located respectively on the paths of one of the first two parts (F ₁₂ ) and one of the last two parts (F2 ₁ ) and introducing between the first two parts (F ₁₁ - F ₁₂ ) on the one hand, and between the last two parts (F ₂₁ -F ₂₂ ) on the other hand , predetermined phase shifts that vary over time, the average values of which, over short periods of time relative to the duration of illumination, are zero. 5. Illuminating device according to claim 3, characterized in that, the initial beam being linearly polarized, the averaging means comprise at least one electro-optical phase shifter element (K) located on the path of one of the two half-beams (F ₂ ) introducing between them a predetermined phase shift that varies over time and whose mean value is zero over a short period of time relative to the lighting duration, and at least two birefringent plates (L. - L ₂ ) located respec tively on the path of one of the first two parts (F ₁₂ ) and one of the last two parts (F ₂₂ ) and creating 90 ° polarization rotations there. 6. Illuminating device intended to provide a coherent illumination beam having a spatial distribution of adjustable intensity, from a source (1) providing an initial beam (F) having a Gaussian intensity distribution (I), characterized in that it comprises means (5 - 6) allowing the transformation of the initial beam (F) into an annular beam, a prism with circular symmetry C, in the shape of a cone whose axis coincides with the axis (z ) of the initial beam (F), means (T) allowing the recombination of the beam thus transformed in a predetermined plane and means (D) allowing the averaging of the beam interference at the output of the recombination means. 7. Illuminating device according to claim 6, characterized in that the recombination means comprise a frustoconical concave mirror (T), the axis of which coincides with the axis of the initial beam. 8. Illuminator device according to any one of claims 6 or 7, characterized in that the averaging means comprise a rotating diffuser (D) placed on the path of the recombined beam. 9. A pattern transfer system making it possible to reproduce the image of an object with non-uniform transparency on a substrate covered with a photosensitive layer, this system comprising an illuminating device providing a beam of illumination of the object, and a objective forming the image of the object on the substrate, the system being characterized in that the illuminating device is according to any one of the preceding claims.

Images